nwcomponents
Network components and the largest component
Syntax
nwcomponents
[netlist]
[, lgc
generate(newvarname)
replace
silent
listall]
generate(newvarname) |
Required. Name of the Stata variable that stores information about components (component membership, or largest-component membership if lgc is given) |
replace |
Replace existing variable |
lgc |
Calculate membership to largest component |
silent |
Suppress display of results |
listall |
Print the full table of components; by default at most the first 30 rows are printed, with a note. Results in r() are unchanged |
Description
Calculate the components of a network or a list of networks. A component is a set of nodes that are only connected among each other. All calculations are performed on the undirected network. Nodes can only belong to one component.
generate() is required and names the new variable that stores the component membership. When option lgc is specified, the command instead stores in that variable information about membership to the largest component.
Examples
. nwwebuse florentine, nwclear
. nwcomponents flomarriage, generate(_component)
----------------------------------------
Network name: flomarriage
Components: 2
_component | Freq. Percent Cum.
------------+-----------------------------------
1 | 15 93.75 93.75
2 | 1 6.25 100.00
------------+-----------------------------------
Total | 16 100.00
This shows that there are two components in the Florentine marriage network. All except one node belong to the first component. Some alternative ways how the commands can be used.
. nwwebuse glasgow
. nwcomponents glasgow1, generate(mycomponent)
. nwcomponents _all, lgc generate(_lgc)
. nwcomponents _all, lgc generate(mylgc)
Supported network types
Binary: yes (only) - component membership is a structural (weak-connectivity) property, tie values are ignored. Directed: yes - uses weak connectivity (ignores tie direction), matching the standard convention for “components” on a directed graph. Weighted: not applicable. Signed: not applicable. Two-mode: same as for the bipartite graph read as a one-mode network (valid; no mode-specific normalisation).
Stored results
Scalars
- r(components) number of components
Matrices
- r(comp_sizeid) distribution over components
See also
nwcommands version 3.17.6, 9 November 2026 (check your installed version)